sendnrw/windows_exporter
2
0
Fork 0
mirror of https://github.com/prometheus-community/windows_exporter.git synced 2026-10-17 10:09:05 +02:00
Code Issues Packages Projects Releases Wiki Activity

hyperv: add the vm_state sub-collector for VMs that are not running (#2765)

Browse Source 
Signed-off-by: Jan-Otto Kröpke <mail@jkroepke.de>
This commit is contained in:
Jan-Otto Kröpke
2026-10-16 21:21:21 +02:00
committed by GitHub
parent ea6126fb67
commit 4f3152d84d
7 changed files with 460 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/collector.hyperv.md
+49 -1
View File
@@ -12,7 +12,7 @@ The hyperv collector exposes metrics about the Hyper-V hypervisor
### `--collector.hyperv.enabled`
Comma-separated list of collectors to use, for example:
`--collector.hyperv.enabled=datastore,dynamic_memory_balancer,dynamic_memory_vm,host,hypervisor_logical_processor,hypervisor_root_partition,hypervisor_root_virtual_processor,hypervisor_virtual_processor,legacy_network_adapter,replica_vm,virtual_hard_disk,virtual_machine_health_summary,virtual_machine_vid_partition,virtual_network_adapter,virtual_network_adapter_drop_reasons,virtual_smb,virtual_storage_device,virtual_switch,wmi_health`
`--collector.hyperv.enabled=datastore,dynamic_memory_balancer,dynamic_memory_vm,host,hypervisor_logical_processor,hypervisor_root_partition,hypervisor_root_virtual_processor,hypervisor_virtual_processor,legacy_network_adapter,replica_vm,virtual_hard_disk,virtual_machine_health_summary,virtual_machine_vid_partition,virtual_network_adapter,virtual_network_adapter_drop_reasons,virtual_smb,virtual_storage_device,virtual_switch,vm_state,wmi_health`
### `--collector.hyperv.virtual-hard-disk.refresh-interval`
Interval at which the `virtual_hard_disk` sub-collector reads the disks in the background, for example `10m`.
@@ -390,6 +390,54 @@ rate(windows_hyperv_virtual_storage_device_io_latency_seconds_total[5m])
| `windows_hyperv_vid_remote_physical_pages` | The number of physical pages not allocated from the preferred NUMA node | gauge | `vm` |
### Hyper-V VM State
Sub-collector `vm_state`, enabled by default. The performance counter objects only have instances for running
virtual machines, so a VM that is off, saved or paused is missing from all other metrics. This sub-collector lists
every virtual machine registered on the host with one WQL query per scrape
(`SELECT Name, ElementName, EnabledState FROM Msvm_ComputerSystem`) against the `root/virtualization/v2` namespace.
The query times out after 5 seconds; a failed or timed-out query fails the `hyperv` collector
(`windows_exporter_collector_success{collector="hyperv"}` is `0`), like the `replica_vm` sub-collector.
On a failover cluster, each node reports the virtual machines it currently owns.
| Name | Description | Type | Labels |
|---------------------------|-----------------------------------------------------------------------|-------|------------------------|
| `windows_hyperv_vm_state` | The state of the virtual machine, 1 if the current state, 0 otherwise | gauge | `vm`, `vm_id`, `state` |
Every virtual machine has one series per `state`: `running`, `off`, `starting`, `stopping`, `saving`, `saved`,
`pausing`, `paused`, `resuming`, `reset` and `unknown`. The value is 1 for the current state and 0 for the others.
`state` is read from `Msvm_ComputerSystem.EnabledState`. A fast-saved VM is reported as `saved`, a VM in a
critical state (for example `RunningCritical`) as its base state, and a value without a mapping as `unknown`.
Replica VMs on a Hyper-V Replica server are turned off and show as `state="off"`; use the
`windows_hyperv_replica_vm_*` metrics to tell them apart.
`vm` is the name of the virtual machine and `vm_id` its GUID, the same keys as the `windows_hyperv_replica_vm_*`
and `windows_hyperv_virtual_hard_disk_*` metrics. VM names don't have to be unique, `vm_id` keeps the series apart.
Virtual machines that are turned off:
```promql
windows_hyperv_vm_state{state="off"} == 1
```
Number of virtual machines per state and host:
```promql
sum by (instance, state) (windows_hyperv_vm_state)
```
Example alert for a virtual machine that is expected to run:
```yaml
- alert: HyperVVMNotRunning
expr: windows_hyperv_vm_state{vm="web01", state="running"} == 0
for: 10m
annotations:
summary: Hyper-V VM {{ $labels.vm }} is not running on {{ $labels.instance }}
```
### Hyper-V WMI Health
Sub-collector `wmi_health`, enabled by default. It runs a minimal WQL query
docs/upgrade/0.32.md
+4 -4
View File
@@ -312,10 +312,10 @@ See [the logical_disk documentation](../collector.logical_disk.md#bitlocker-stat
### New default sub-collectors
| Collector | New sub-collectors enabled by default |
|-------------|---------------------------------------------------------|
| `hyperv` | `host`, `replica_vm`, `virtual_hard_disk`, `wmi_health` |
| `mscluster` | `shared_volumes` |
| Collector | New sub-collectors enabled by default |
|-------------|---------------------------------------------------------------------|
| `hyperv` | `host`, `replica_vm`, `virtual_hard_disk`, `vm_state`, `wmi_health` |
| `mscluster` | `shared_volumes` |
These add new metrics. To keep the 0.31 set, list the sub-collectors explicitly with `--collector.hyperv.enabled` or `--collector.mscluster.enabled`.